Difference between explain analyze and real execution time - Mailing list pgsql-performance

From Artur Zając
Subject Difference between explain analyze and real execution time
Date
Msg-id 000001cb849e$1dce2720$596a7560$@ang.com.pl
Whole thread Raw
Responses Re: Difference between explain analyze and real execution time  (Tom Lane <tgl@sss.pgh.pa.us>)
Re: Difference between explain analyze and real execution time  (Robert Haas <robertmhaas@gmail.com>)
List pgsql-performance

I have some simple query (executed with time command):

 

 

time psql  -c 'explain analyze SELECT te.idt FROM t_positions AS te  JOIN t_st AS stm ON (te.idt=stm.idt AND 4=stm.idm)   WHERE te.idtr IN (347186)'

 

 

                                                                       QUERY PLAN

--------------------------------------------------------------------------------------------------------------------------------------------------------

Nested Loop  (cost=0.00..33.33 rows=2 width=4) (actual time=0.297..0.418 rows=3 loops=1)

   ->  Index Scan using t_positions_index1 on t_positions te  (cost=0.00..8.43 rows=3 width=4) (actual time=0.140..0.148 rows=3 loops=1)

         Index Cond: (idtr = 347186)

   ->  Index Scan using t_st_index4 on t_st stm  (cost=0.00..8.29 rows=1 width=4) (actual time=0.078..0.079 rows=1 loops=3)

         Index Cond: ((stm.idt = te.idt) AND (4 = stm.idm))

Total runtime: 0.710 ms

(6 rows)

 

 

real    0m3.309s

user    0m0.002s

sys     0m0.002s

 

Why there is so big difference between explain analyze (0.710 ms) and real execution time (3309 ms)? Any suggestions?

 

Psql only execution time:

 

time psql -c 'explain analyze SELECT blabla()'

ERROR:  function blabla() does not exist

 

real    0m0.011s

user    0m0.001s

sys     0m0.004s

 

SELECT version();

 

PostgreSQL 9.0.1 on i686-pc-linux-gnu, compiled by GCC gcc (GCC) 4.1.2 20070502 (Red Hat 4.1.2-12), 32-bit

 

-------------------------------------------

Artur Zajac

 

 

pgsql-performance by date:

Previous
From: Mladen Gogala
Date:
Subject: Re: MVCC performance issue
Next
From: Robert Haas
Date:
Subject: Re: Running PostgreSQL as fast as possible no matter the consequences